Opsgenie

Opsgenie is Atlassian's on-call and incident management platform, widely deployed in enterprise SRE and DevOps teams. Its REST API enables agents to create alerts, query on-call schedules, acknowledge and close alerts, and manage escalation policies programmatically. For agents that detect system anomalies or operational issues, Opsgenie's alert creation API is the direct path to notifying the right on-call engineer through configured escalation policies — without requiring the agent to know who is on-call or how to reach them.

Authentication uses API keys that can be team-scoped or integration-scoped. Opsgenie supports both API key and OAuth authentication. Integration-scoped API keys (created as Opsgenie integrations) are the appropriate mechanism for agent alert creation — they appear as a named integration in the Opsgenie audit trail and can be scoped to specific teams.

Documentation is thorough and covers the API with clear examples. The alert attribute reference is particularly useful for agents that need to populate rich context in created alerts. Teams integrating agents with Opsgenie for automated incident creation will find the documentation covers the alert lifecycle, routing logic, and API surface comprehensively.

The API covers alerts, incidents, on-call schedules, teams, and services. The alert creation API is the primary agent integration surface — agents post alerts with relevant context (source, message, priority, tags, additional details) and Opsgenie routes them through configured policies to the appropriate responders. The on-call schedule API enables agents to query who is currently on-call for informational purposes or for building acknowledgment workflows.

Reliability is enterprise-grade for on-call infrastructure. Opsgenie maintains the availability and delivery guarantees required for incident notification — alert delivery is the core reliability contract, and Opsgenie's multi-channel delivery (push, SMS, voice) ensures notifications reach responders even if individual channels are unavailable. Agents creating alerts for critical incidents should implement retry logic to ensure delivery.